Work has begun on the cavity that will be used to lock the two lasers required for the STIRAP optical transfer. The mirrors for the cavity consist of one plane mirror and one spherical mirror with a radius of curvature of 50cm. The reflectivity of both mirrors is >99.91% which allows a cavity with a finesse of ~3000 to be made.

The blue line in the figure shows the signal from a commercial etalon (Coherent, 33-6230-001) with a 300 MHz free spectral range (FSR). The red line shows the signal from a home made cavity consisting of the plane and spherical mirrors separated by ~20cm. Using the signal from the commercial etalon the FSR of the home made cavity was measured to be 768(5)M MHz. This corresponds to a calculated mirror separation of 19.5(1) cm in good agreement with what we measure in the lab.

The next step is to secure these mirrors onto either end of a 20 cm long Zerodure spacer. A PZT will be placed between one mirror and the spacer to allow the cavity to be locked to a Pound-Drever-Hall signal from a frequency stabilised 852 nm laser.

We have located two regions where the evaporation efficiency ~3 (purple dashed boxes), here subsequent evaporation by titling the trap with the magnetic field gradient can lead to the production of 133Cs condensates containing up 8x10⁴ atoms. One region is the standard window used to produce 133Cs BEC located between 20-25G. The second window spans 18G to 19.6G and marks the discovery of a new bias field region where Cs BEC can be achieved.

To align the crossed dipole trap both beams were aligned independently using the single beam configuration. As this trap can only confine low field seeking states atoms in this trap are prone to Majorana losses if the dipole beam is aligned through the magnetic field zero. This loss was used to locate the magnetic field zero in the vertical and horizontal directions as shown in the figure. The two data sets shown correspond to the alignment of each dipole beam with a power of 6 W. The fits to these data are to guide the eye.

Once the magnetic field zero was located each dipole beam was positioned 200 µm below the magnetic field zero. Bose-Einstein condensates consisting of 2x10⁶ 87Rb atoms can now be made in either single beam trap while 87Rb condensates of 2x10⁶ atoms and 133Cs condensates consisting of 8x10⁴ atoms can now be made in the crossed dipole trap.

For details on the experimental procedures followed to reach these results please see our recent publications.

Phys. Rev. A 84 011603 (2011); arXiv:1102.1576

We report the formation of a dual-species Bose-Einstein condensate of 87Rb and 133Cs in the same trapping potential. Our method exploits the efficient sympathetic cooling of 133Cs via elastic collisions with 87Rb, initially in a magnetic quadrupole trap and subsequently in a levitated optical trap. The two condensates each contain up to 2x10⁴ atoms and exhibit a striking phase separation, revealing the mixture to be immiscible due to strong repulsive interspecies interactions. Sacrificing all the 87Rb during the cooling, we create single-species 133Cs condensates of up to 6x10⁴ atoms.

For more information see Publications.

Details on this significant result can be found in one of our recent publications Phys. Rev. A 84 011603 (2011).

This has had a marked effect on sympathetic cooling of 133Cs, which we are now able to cool with only a small deviation of the 87Rb number from the single species case. After the RF evaporation we obtain a similar number of Cs atoms, at a phase-space density which is an order of magnitude higher than previously.

The figure highlights the atom number remaining in the trap at different temperatures within the RF evaporation sequence. Green squares highlight 133Cs data points, solid red circles highlight 87Rb cooling Cs data, and hollow purple circles correspond to the single species 87Rb case. The grey data points highlight previous data. All lines have been added to guide the eye. Notice how the 87Rb cooling Cs case differs little from the single species case, whilst previously there was a drastic difference. It can also be observed that it is possible to cool the 133Cs to lower temperatures now.

The second paper is concerned with the condensation of 133Cs via sympathetic cooling with 87Rb. We then report the production of simultaneous 87Rb and 133Cs BECs and the formation of Cs2 molecules.

The figure, taken from the 87Rb BEC paper, highlights the sympathetic cooling of the m_F=-1 and m_F=0 states of 87Rb, after rethermalising elastic collisions with atoms in the m_F=+1 state. Figure (a) demonstrates how the temperatures of the 0 and -1 states follow that of the +1 state after collisions. The inset presents the potential for each state. Here solid lines correspond to the trap depths, the dashed line corresponds to the trap depth along the beam and points correspond to temperatures from measurements. Figure (b) demonstrates how well sympathetic cooling works, as only the +1 atoms are lost initially, due to efficient sympathetic cooling. Once all +1 atoms are lost the m_F=0 atoms take over as the refrigerant and cool the -1 atoms. Solid lines are added to guide the eye.

To read the abstracts, see Publications.

We can create Cs Feshbach molecules by applying a magnetic field sweep across a Feshbach resonance to a high phase-space density sample of Cs atoms. We use a Feshbach resonance located at 19.8G and observe molecule production from either a Cs Bose-Einstein condensate or a high phase-space density thermal cloud. The molecular sample is distinguished from the atom cloud via spatial Stern-Gerlach separation. In the image a magnetic field gradient levitates the atom cloud while the molecules accelerate downward at approximately 0.6g.

To image the molecular cloud a reverse field sweep across the Feshbach resonance is applied. This brings the molecules above the scattering continuum and results in them quickly dissociating back into free atoms. An absorption image of the reconverted atoms then reveals the spatial distribution of the molecules. By levitating the molecules we are able to measure the molecular magnetic moment. We find that µ = 0.92(1) µ_B, which is in good agreement with previous data.

September 2008: Our recent work involving modulation transfer spectroscopy has now been published in Measurement Science and Technology, Meas. Sci. Technol. 19, 105601 (2008). This work demonstrates that both the gradient and amplitude of modulation transfer spectroscopy signals, for the 87Rb F = 2 to F' = 3 and the 85Rb F = 3 to F' = 4 transitions, can be significantly enhanced by expanding the beams, improving the signals for laser frequency stabilization. The modulation transfer signal for the 85Rb F = 2 to F' transitions is also presented to highlight how this technique can generate a single, clear line for laser frequency stabilization even in cases where there are a number of closely spaced hyperfine transitions.

The figure shows a comparison between (a) modulation transfer spectroscopy and (b) frequency modulation (FM) spectroscopy for the 87Rb F = 2 to F' and 85Rb F = 3 to F' transitions at a modulation frequency of 12.35 MHz. The reference saturation absorption/ hyperfine pumping spectroscopy signal is shown in (c). This figure shows that, unlike FM spectroscopy, modulation transfer spectroscopy has one large zero crossing corresponding to the closed transition and a flat background signal.